Re: Basic question?What 's the key if there 's no FD(Functional Dependencies)?

NENASHI, Tegiri a écrit :

> Cimode wrote:
> > A second...
> >
> > Let R1 relation of RealPeople, R2 relation of Impostors, R3 relation of
> > Impostors and RealPeople expressed as R3 = R1 UNION R2
> > R1 and R2 are disjoint therefore:
> > R1 INTERSECT R2 = Empty Set
> > R1 MINUS R2 = R1
> > R2 MINUS R1 = R2
> >
> > implies...
> >
> > R1 MINUS R2
> > UNION
> > R2 MINUS R1
> > UNION
> > R1 INTERSECT R2
> >
> > is the same as
> >
> > R1 UNION R2 therefore R3 (call it AllPeople). In english ? Can
> > relation AllPeople defined as the union of Impostors AND RealPeople be
> > updated? Answer YES. Is it updated as an Impostor NO Is it updated as
> > RealPeople NO. It is simply updated as AllPeople. Period.
> >
> > Neither R1 nor R2 are updated in fact. A major advantage of RM is the
> > ability to work in closure with sets (meaning independently from values
> > included in the sets at the first place)
> >
> > For instance,
> >
> > R1:{A, B, C}
> > R2:{D, E, F}
> >
> > R3 = R1 UNION R2 --> leads to R3 :{A, B, C, D, E, F}
> > If inserting R4: {G}
> > Therefore {A, B, C, D, E, F} UNION {G} you produce R5 ={A, B, C, D, E,
> > F, G}
> > Neither R1, R2, R4 are updated in the process just R3 and R5 are
> > exrpressed
> >
> > A consequence of assimiliating relations to revalues sets instead of
> > focusing on their transformational properties is a common error in RM
> > theory. Operations between 2 relations necessarily produce a new
> > relation. Even an update to such relation produces a new relation.
>
> To produce a new relation it is a very good point if the computer is
> named "math". One knows that the "math" has infiite speed and infinite
> memory. Helas, with the real computer, one must think about
> efficiency so your idea will not work in the real world.
Efficiency of any implementation in *da real world* depends on the soundness of the abstract theory that supports it. The answer to some of the questions asked in RM perspective not bulshit SQL biased view of RM.

> Tegi
Received on Fri Nov 10 2006 - 10:28:51 CET